Skip to main content

Table 2 Error rates, types of errors, and potential clinical impact for the preintervention and intervention periods, by study group

From: Effectiveness of a ‘do not interrupt’ vest intervention to reduce medication errors during medication administration: a multicenter cluster randomized controlled trial

 

Experimental group

Control group

 

TOE

Preintervention period (A)

(n = 1235)

Intervention period (B)

(n = 2653)

p-value

A vs. B

(*)

Preintervention period (C)

(n = 1211)

Intervention period (D)

(n = 3373)

p-value

C vs. D

(*)

p-value

B vs. D

(†)

Number (%) of OE with at least one error

61 (4.94)

188 (7.09)

0.013

78 (6.44)

210 (6.23)

0.845

0.355

Total number of errors a

62

191

 

81

213

  

Types of errors, n (%)

 Wrong dosage-form error

18 (29.0)

76 (39.8)

0.170

17 (21)

84 (39.4)

0.004

0.851

 Unauthorized drug error

9 (14.5)

34 (17.8)

0.686

5 (6.2)

28 (13.1)

0.138

 Omission error

4 (6.5)

7 (3.7)

0.471

1 (1.2)

1 (0.5)

0.476

 Wrong-dose error

30 (48.4)

67 (35.1)

0.085

47 (58.0)

79 (37.1)

0.001

0.046

 Wrong administration technique error

0 (0.0)

7 (3.7)

0.199

10 (12.3)

14 (6.6)

0.169

 Wrong drug-preparation error

1 (1.6)

0 (0)

0.245

1 (1.2)

7 (3.3)

0.452

Potential clinical impact of errors, n (%)

 No clinical impact

43 (69.4)

130 (68.1)

0.974

56 (69.1)

171 (80.3)

0.060

 Serious or significant clinical impact b

19 (30.6)

61 (31.9)

 

25 (30.9)

42 (19.7)

  
  1. a One OE can be associated with several errors and different clinical impact
  2. b No OE had a potential life-threatening impact
  3. * p-value corresponds to either a Fisher test or a Chi-2 test
  4. p-value corresponds to the interaction term of the mixed logistic regression model (see statistical methods). Due to a low numbers of OE, the model could not converge for unauthorized drug error, omission error and wrong administration technique and drug-preparation error. No calculation was possible for the p-value of the potential clinical impact because an OE could be associated with several errors with different clinical impact